> ## Content Index > Fetch the complete content index at: https://www.therecursive.com/llms.txt > Use this file to discover other available public pages before exploring further. # Unpacking Model ML's Dubai Plans as GCC Sovereign Funds and Banks Look Beyond General-Purpose LLMs - URL: https://www.therecursive.com/unpacking-model-mls-dubai-plans-as-gcc-sovereign-funds-and-banks-look-beyond-general-purpose-llms/ - Published: 2026-11-12T11:57:27.000Z - Updated: 2026-11-12T11:57:27.000Z - Description: With an Abu Dhabi sovereign wealth fund already among its clients, Model ML has opened a Dubai office to expand across the GCC. Ajinkya Patwardhan explains why banks and investors are moving beyond general-purpose LLMs, how agentic AI is changing finance, and where human judgement remains essential. - Author: Etien Yovchev - Tags: Fintech, Agentic AI, UAE, GCC, AI in finance, banking If there is one thing that attracts finance professionals, it is more hours in the day, says[ Ajinkya Patwardhan](https://www.linkedin.com/in/ampatwardhan1/?ref=therecursive.com), Strategic Account Executive, MENA at [Model ML](https://modelml.com/?ref=therecursive.com). Giving those hours back is Model ML’s business. Model ML is a vertical harness for finance: Agentic AI purpose-built to automate document-heavy work such as pitch books, investment memos and financial models. The company has raised more than $100 million and counts HSBC, Santander, PwC and Deloitte among its customers. At the end of October, Model ML opened its first Middle East office in Dubai. From there, it plans to serve banks, sovereign wealth funds and family offices across the GCC. Its Gulf clients already include an Abu Dhabi sovereign wealth fund and the investment bank Awad Capital, according to the company. The Recursive co-founder Etien Yovchev spoke with Patwardhan at[ Dubai FinTech Summit](https://dubaifintechsummit.com/?ref=therecursive.com), held on 2 and 3 November. They discussed what opportunity Model ML sees in the Gulf, where AI stops and human judgement starts in investing, and how the company plans to grow in the region. ![](https://storage.ghost.io/c/b8/ae/b8aee878-7ddc-40ad-8f68-92a28c560560/content/images/2026/11/1793702126315.jpg) ## **Model ML has just opened its first Middle East office in Dubai. What opportunity do you see in the GCC?** We think the GCC is always on the cutting edge of AI adoption. There’s a lot of interest in how enterprises can use AI to become more productive and efficient, within the conditions that matter most to this market: information security, data residency and effective use of their people. Within that context, we see many organisations willing to embrace and adapt to the new reality of working with AI. There’s a huge willingness to run pilot programmes and test the technology in small factories of innovation inside banks, consulting firms and other institutions. We’re also noticing that the region doesn’t only adopt AI. It builds with it. In several of the organisations we serve, AI fluency is very high. There are AI champions who are power users of frontier lab models, and people who experiment and build AI tooling themselves. The young workforce here is AI native. It has already absorbed how this technology can make people more competitive inside their organisation, and their organisation more competitive in its market. ## **What has changed in the past one to two years to move AI in finance from experimentation into core operations?** The technology has basically proven itself. Twelve to fourteen months ago, you were struggling to explain to clients why a vertical harness is distinct from a frontier lab, or what a vertical harness even is. You had to explain how agentic workflows happen and how you embed them in an organisation. Now people are actively going out to the market and issuing RFPs for vertical harnesses, because they’re already sold on the utility of the technology. That transition is a function of the good work people in our business have done. It also reflects how the economics of using AI have shifted in the last 12 to 18 months, and how much better open-weight models have become. And people have realised that a general-purpose model is not nearly as effective as a deep subject-matter specialist and a team building for a specific use case. ## **You help investment teams reach the productivity of teams several times their size. How does your AI do that, and what is its advantage over general LLMs?** The investment process involves a lot of manual document creation, document review, analysis and presentation. All of it goes into building a thesis that eventually reaches the investment committee. LLMs are fundamentally good at acting as subject-matter experts in document review, document creation, presentation and analysis. Purpose-built vertical harnesses are even better at it within one vertical, which for us is finance. That means you can massively cut the time it takes to put together an investment deck or memorandum, or to build an LBO or DCF model. The same goes for responding to an investment prospectus from a third-party fund. These tasks are common across investing. They’re operationally heavy and consume a lot of time and energy. Giving that time back lets investors stress-test the quality of a counterpart if they’re a fund-of-funds investor, think deeply about the industry they invest in, or develop their own taste and judgement. That is more valuable than stressing about the third footnote on slide 68\. Model ML takes care of that, so you can focus on the more value-added parts of your job. ## **How do you expect AI to change the way investment teams work over the next three to five years? What will it take over, and where will humans remain essential?** You’re already seeing AI take on end-to-end workflows: document creation, pitch books, building a confidential information memorandum, even the first-review draft of an investment memo. A lot of the grunt work needed to produce material robust enough for the investment committee, and the analysis behind it, AI can already do. That will only accelerate. What’s important is the human-in-the-loop element of an agentic workflow. For an investor, being the human in the loop is not just signing off on something or clicking a final button of approval. It means applying your proprietary taste, your judgement, your relationships and your in-depth expertise in a particular market or sector to give more colour to an agent. Take a simple example. Agents are really good at reading earnings reports. You can have one digest a report and produce all kinds of analysis to help form an investment opinion. But an agent can’t yet pick up the phone, call a banker who covers that institution and talk about what they’re seeing on the ground. So there will always be an element of human taste, judgement and involvement in the investment process. But a lot of the drudge work will be shortcut through the use of AI. ## **What is your go-to-market (GTM) and growth strategy for the MENA region over the next year?** In many ways it will mirror what we’ve seen in other regions, with a flavour that is unique and respectful to the priorities of institutions here. In other regions we led with banks and finance, then expanded to consultancies and asset managers. We’re live with three of the Big Four. So on the one hand, it’s about localising relationships that were built globally. The local organisations and member firms of institutions we already serve should get the same high-quality product we provide elsewhere. On the other hand, it’s about making sure local firms here, including sovereigns, local banks and local asset managers, get access to the same productivity improvement, so they can compete on a level playing field. I see our GTM strategy evolving through word of mouth. Once we get the product into enough hands, you get a buzz around how competitive and useful it is. If there’s one thing that’s attractive to finance professionals, it’s more hours in the day, whether to do more work, close more deals and chase net new revenue, or to get a few extra hours of sleep. We’re quite confident that if people try the product, we’ll get a lot of granular word-of-mouth advertising. And working with publications like yours, taking part in conferences and getting the word out about what Model ML has built will be central to our go-to-market strategy.